Computerized system and method for dynamic camera and scene adjustment for autonomous charging
The disclosed systems and methods provide a novel framework that provides mechanisms for a hands-free, autonomous electrical connection of an electric charger to an electric vehicle (EV), and subsequent charging. The disclosed framework utilizes an automated connection device (ACD) as an intermediary between the charger and the EV. The ACD is configured for automatically determining a precise location of the charging inlet on the EV and then automatically establishing an electrical connection with the EV so that the EV can receive a charge. The ACD performs the disclosed precise positional and directional navigation to the EV inlet based on deep neural network analysis of captured imagery of the inlet. In some embodiments, the images can be modified so as to highlight and/or assist the ACD's navigation towards to the inlet in order to maximize invariance.
1. A method comprising:
identifying, by an automated connection device (ACD), a first set of images, each image comprising content representing a charging port on an electric vehicle (EV);
analyzing, by the ACD, each image in the first set of images, and identifying information related to specific portions of the charging port;
determining, by the ACD, a region of interest (ROI) based on the identified information in the first set of images;
adjusting, by the ACD, parameters of a camera that control how images are captured, the adjustment of the parameters being based on the determined ROI in the first set of images,
wherein the adjusted parameters comprise information related to automatically adjusting auto-focus features of the camera;
determining, by the ACD, navigation information for an arm of the ACD respective to the charging port of the EV based on usage of the adjusted parameters by the camera; and
automatically establishing, by the ACD based on the navigation information, a physical connection with the arm of the ACD and the charging port of the EV, the physical connection comprising a plug at a distal end of the arm being automatically inserted into the charging port.
2. The method of claim 1 , further comprising:
charging, via the ACD, the EV via the established physical connection, wherein a charge is provided by a charging source that is associated with the ACD.
3. The method of claim 1 , wherein the determination of the navigation information comprises:
capturing, by the camera associated with the ACD, a second set of images based on the adjusted parameters; and
analyzing, by the ACD executing a Convolutional Neural Network (CNN), the second set of images, wherein the navigation information is based on the analysis of the second set of images.
4. The method of claim 3 , further comprising:
determining a set of bounding boxes, masks, or a combination thereof based on the ROI, each bounding box corresponding to a portion of the charging port, wherein the bounding boxes are applied during the capturing the second set of images.
5. The method of claim 3 , wherein the adjustment of the camera parameters occurs during the capturing of the second set of images.
6. The method of claim 1 , wherein the camera parameters comprise at least one of exposure, backlight compensation, focus, sharpness, contrast, saturation, brightness, illumination intensity and vectored illumination.
7. The method of claim 1 , wherein the navigation information comprises information related to a direction, distance and/or trajectory the arm of the ACD is required to move to establish the physical connection.
8. The method of claim 1 , wherein the analysis of the first set of images and the determination of the ROI is based on the ACD executing a Convolutional Neural Network (CNN).
9. The method of claim 1 , further comprising:
receiving a request to charge the EV, the request based on an arrival of the EV at a charging station, wherein the first set of images are identified based on the request.
10. The method of claim 1 , wherein the identifying of the first set of images comprises the camera capturing the first set of images.
11. An automated connection device (ACD) comprising:
a camera; and
a processor, the processor configured to:
identify a first set of images, each image comprising content representing a charging port on an electric vehicle (EV);
analyze each image in the first set of images, and identify information related to specific portions of the charging port;
determine a region of interest (ROI) based on the identified information in the first set of images;
adjust parameters of the camera that control how images are captured, the adjustment of the parameters being based on the determined ROI in the first set of images,
wherein the adjusted parameters comprise information related to automatically adjusting auto-focus features of the camera;
determine navigation information for an arm of the ACD respective to the charging port of the EV based on usage of the adjusted parameters by the camera; and
automatically establish, based on the navigation information, a physical connection with the arm of the ACD and the charging port of the EV, the physical connection comprising a plug at a distal end of the arm being automatically inserted into the charging port.
12. The ACD of claim 11 , wherein the processor is further configured to:
charge the EV via the established physical connection, wherein a charge is provided by a charging source that is associated with the ACD.
13. The ACD of claim 11 , wherein the processor is further configured to:
capture a second set of images based on the adjusted parameters; and
analyze, via execution of a Convolutional Neural Network (CNN), the second set of images, wherein the navigation information is based on the analysis of the second set of images.
14. The ACD of claim 13 , wherein the processor is further configured to:
determine a set of bounding boxes, masks, or a combination thereof based on the ROI, each bounding box corresponding to a portion of the charging port, wherein the bounding boxes are applied during the capturing the second set of images.
15. The ACD of claim 11 , wherein the navigation information comprises information related to a direction, distance and/or trajectory the arm of the ACD is required to move to establish the physical connection.
16. The ACD of claim 11 , wherein the analysis of the first set of images and the determination of the ROI is based on execution of a Convolutional Neural Network (CNN).
17. A non-transitory computer-readable storage medium tangibly encoded with computer-executable instructions, that when executed by an automated connection device (ACD), performs a method comprising:
identifying, by the ACD, a first set of images, each image comprising content representing a charging port on an electric vehicle (EV);
analyzing, by the ACD, each image in the first set of images, and identifying information related to specific portions of the charging port;
determining, by the ACD, a region of interest (ROI) based on the identified information in the first set of images;
adjusting, by the ACD, parameters of an associated camera that control how images are captured, the adjustment of the parameters being based on the determined ROI in the first set of images,
wherein the adjusted parameters comprise information related to automatically adjusting auto-focus features of the camera;
determining, by the ACD, navigation information for an arm of the ACD respective to the charging port of the EV based on usage of the adjusted parameters by the camera; and
automatically establishing, by the ACD based on the navigation information, a physical connection with the arm of the ACD and the charging port of the EV, the physical connection comprising a plug at a distal end of the arm being automatically inserted into the charging port.
18. The non-transitory computer-readable storage medium of claim 17 , wherein the determination of the navigation information comprises:
capturing, by the camera associated with the ACD, a second set of images based on the adjusted parameters; and
analyzing, by the ACD executing a Convolutional Neural Network (CNN), the second set of images, wherein the navigation information is based on the analysis of the second set of images.